Hybrid Hierarchical Backtracking Search Optimization Algorithm and Its Application

Abstract

As a young intelligence optimization algorithm, backtracking search optimization algorithm (BSA) has been used to solve many optimization problems successfully. However, BSA has some disadvantages such as being easy to fall into local optimum, lacking the learning from the optimal individual, and being difficult to adjust the control parameter F. Motivated by these analyses, to improve the optimization performance of the original BSA, a new hybrid hierarchical backtracking search optimization algorithm (HHBSA) is proposed in this paper. In the proposed method, a two-layer hierarchy structure of population and a randomized regrouping strategy are introduced in the proposed HHBSA for improving the diversity of population, a mutation strategy is used to help the population when the evolution is stagnant and an adaptive control parameter is presented to increase the learning ability of the BSA. To verify the performance of the proposed approaches, 48 benchmark functions and three real-world optimization problems are evaluated to test the performance of the proposed approach. Experiment results indicate that HHBSA is competitive to some existing EAs.